Memory: Hardware and Allocation Summer 2011 Cornell University 1 - - PowerPoint PPT Presentation

1

CS 4410 Operating Systems

Memory: Hardware and Allocation

Summer 2011 Cornell University

2

Today

• How the memory is shared among the ready

processes?

• Memory
• Address protection
• Logical vs Physical Address
• Contiguous memory allocation

3

Storage Hierarchy

Registers Cache (L1, L2, L3) Memory Hard Disk

4

Memory

• A large array of words.
• Word = 4 or 8 bytes.
• One address for every

word.

• Content:
• Instructions
• Data

5

Instruction execution cycle

• Fetch instruction from memory.
• The PC saves its address.
• Decode instruction.
• Fetch operants form memory.
• Execute the instruction.
• Store result in memory.
• Program and data should be in Memory to

become useful.

6

Memory Management

• Is memory shared between processes? How?
• Monoprogramming
• Only one process is ready and loaded into memory.
• It shares the memory space with the OS.
• Is it efficient?
• Multiprogramming
• Fixed or variable partitions for every ready process.
• 2 problems: relocation, protection. Solutions?
• Timesharing
• Swapping

–

Entire process (code, data) is transferred from disk to memory, and vice versa.

• Virtual Memory

–

Processes can run when they are partially in the memory.

7

Memory Management

• What about the memory addresses?
• Monoprogramming

– The physical addresses are known to the programmer.

• Multiprogramming

– The physical addresses are known at the loading time.

• Timesharing

– The physical addresses are known at the execution. – The CPU understands logical addresses. – The Memory understands physical addresses.

8

Memory Management

• Basic concerns:
• Allocation
• Relocation
• Protection

9

Hardware address protection

• Multitasking OS
• Multiple processes loaded in the memory.
• Each process has a separate memory space.
• HW+OS are responsible for address protection.

OS process process process

25600 30004 42094 88000 30004 12090

base limit

10

Logical vs Physical Address

• Multitasking OS
• Memory management: Swap, Virtual Memory
• Logical Address
• Address generated by the CPU
• Address loaded into PC
• Address used in a program
• Physical Address
• Address seen by the memory unit

11

Address Binding

• Logical → Physical
• Execution time
• Logical Address ↔ Virtual Address
• Memory-Management Unit (MMU)
• Hardware device
• Run-time mapping

CPU memory +

14000 Relocation register Logical address 346 Physical address 14346

12

Dynamic Loading

• Using Virtual memory:
• Better memory-space utilization
• The main program is loaded into memory.
• A routing of the program is not loaded until it is

called.

• It is users' responsibility.

13

Dynamic Linking

• Using Virtual memory:
• Without this, each program should include a

copy of its language library.

• It waists disk and memory space.
• With Dynamic Linking:
• A stub substitutes a library-routine reference.
• When stub is executed:

– It checks if the routine is in the memory. – If not, the program loads the routine.

14

Contiguous Memory Allocation

• Share memory between OS and multiple processes.
• Each process is contained in a single contiguous section in memory.
• Memory protection:
• CPU scheduler selects process for execution.
• The dispatcher loads the relocation and limit registers.

CPU < + memory limit register relocation register trap: addressing error logical address no yes physical address

15

Allocation Strategies

• Fixed-sized partitions
• The degree of multiprogramming is bound by the number of

partitions.

• Variable-partition scheme
• The OS keeps a table indicating which parts of memory are

available and which are occupied.

• The OS tries to fit the memory demands of a process in the

available memory space.

• Dynamic storage allocation problem:

– First fit – Best fit – Worst fit

16

Fragmentation

• External fragmentation
• First-fit, Best fit
• There is enough total memory space to satisfy a request but

the available spaces are small and not contiguous.

• Solution 1: Break the physical memory into fixed-sized blocks

and allocate memory in units based on block size.

• Internal fragmentation: the allocated memory is slightly

larger than the requested memory.

• Solution 2: Compaction

17

Today

• How the memory is shared among the ready

processes?

• Memory
• Address protection
• Logical vs Physical Address
• Contiguous memory allocation